Predicting cumulative load during running using field-based measures.

Abstract

The main objective was to investigate whether the cumulative load of the lower limbs, defined as the product of external load and step rate, could be predicted using spatiotemporal variables gathered with a commercially available wearable device in running. Therefore, thirty-nine runners performed two running tests at 10 and 12km/h, respectively. Spatiotemporal variables (step rate, ground contact time, and vertical oscillation) were collected using a commercially available wearable device. Kinetic variables, measured with gold standard equipment (motion capture system and instrumented treadmill) and used for the calculation of a set of variables representing cumulative load, were peak vertical ground reaction force (peak vGRF), vertical instantaneous loading rate (VILR), vertical impulse, braking impulse, as well as peak extension moments and angular impulses of the ankle, knee and hip joints. Separate linear mixed-effects models were built to investigate the prediction performance of the spatiotemporal variables for each measure of cumulative load. BMI, speed, and sex were included as covariates. Predictive precision of the models ranged from .11 to .66 (R2 m ) and .22 to .98 (R2 c ), respectively. Greatest predictive performance was obtained for the cumulative peak vGRF (R2 m =.66, R2 c =.97), VILR (R2 m =.43, R2 c =.97), braking impulse (R2 m =.52, R2 c =.98), and peak hip extension moment (R2 m =.54, R2 c =.90). In conclusion, certain variables representing cumulative load of the lower limbs in running can be predicted using spatiotemporal variables gathered with a commercially available wearable device.